USER`S MANUAL John B. Ketterson K.Rivkin
Contents
1. omega Similar to EXEC with one difference obtained results are expanding onto the basis of the eigenmodes should be either calculated via SOLVE_MATRIX or loaded via EIGEN LOAD Proceeds with Runge Kutta integration for parameters set by SET CURRENT SET_HRF etc for the time period Endtime double Besides that the program outputs into the file Filename character time average magnetization along x y and z axis and the system s total energy If r f field is present in the system omega double indicates the r f field s frequency Unlike EXEC the results are projected onto the basis of first Nmodes eigenvectors every Vtime double time Modes dat and phases dat files contain absolute values and phases of such projections More powerfully exited modes are than presented in modej dat files where j is a 24 mode number from low to high frequencies and collectively in modes01 dat and modes03 dat those with maximum excitation higher than 10 and 30 of the maximum excitation observed e Displaying the eigenmodes As a result of solving the eigenvalue equation we have a set of modes each of them represented by 2 complex numbers per lattice site These 4 numbers uniquely define the oscillation of magnetic moments in the plane perpendicular to the direction of equilibrium magnetization In general such oscillation orbits are ellipses Here you can see that the ellipse can be uniquely identified by another 4 param2. 2C e Linear field Type 2 is defined by the same set of components and in addition to them also by an integer i which defines the coordinate upon which the r f field linearly depends 1 for x 2 for y 3 for z As an example the field that linearly depends on x i 1 will look like h Re amp C 9C 2C ax bye where a and b are selected in such a way that the field is 0 in the middle of the sample positive on one side and negative on the other side 9 If a file is being read Type 3 than 12 integer defines the file s type currently only 12 1 text file is being supported and Filename character is the file s name Text file has a format x y z real cx imag cx real cy imag cy real cz imag cz where x y z are integers and the rest are real numbers corresponding to real and imaginary parts of parameters cx cy CZ R f spin polarized current Type 4 can be defined by its magnitude icur double and the coordinate components of its polarization cx cy and cz double SET_VRML Sevx Scvy Sevz Vtype Sets the parameters of VRML output Scvx Scvy Scvz integers define over how many elements along x y and z coordinates respectively the averaging should be performed with the result being used in making a vrml file Vtype integer defines whether the elements should be presented as they are Vtype 1 or as slices alongside the z direction Vtype 1 By default Sevx Scvy Scvz Vtype
3. Type integer is 1 for the text 2 for binary mode 3 for VRML Filename character is the file s name Using text files involves certain loss of precision Analogously magnetic fields can be loaded using 20 H_LOAD Type Filename Type integer is 1 for the text 2 for the binary mode Filename character is the file s name H_AVERAGE Cxl Cyl Czl Cx2 Cy2 Cz2 Averages the field in occupied sites in a rectangular prism with the lower left corner with coordinates Cx1 Cyl Cz1 double and the upper right corner with coordinates Cx2 Cy2 Cz2 double The resulting field is displayed on the screen SET_TIMESTEP Timestep If one wishes to use Runge Kutta integration in particular the next command EXEC one first has to set the timestep This commands can set the timestep explicitly to the value given by Timestep double in the units defined by SET GYRO or if timestep lt 0 it will run a program that finds the optimal value of timestep automatically The method used to calculate the optimal value is overly conservative in most cases it returns somewhat smaller timestep value than the ones traditionally used EXEC Endtime Vtime Filename omega Proceeds with Runge Kutta integration for parameters set by SET CURRENT SET_HRF etc for the time period Endtime double If Vtime double is greater than zero every Vtime interval the magnetization distribution is being saved into VRML file therefore the e
4. depict the orbits of each of the spins with the principal radii proportional to the precession amplitude The color also signifies the amplitude of precession red for the maximum amplitude yellow for the amplitudes above 50 of this maximum amplitude and blue for the lower amplitudes The white lines emanating from the orbit centers identify the relative phases of the oscillations the orbits shown here are actually located in different coordinate systems defined by the local direction of magnetization as noted above Example a o0 0000000 o o 6 GG OG Oe e 000600e 60 05000060906000 0 G 6 00000000 e 5 00 46 09 04 o e o w000 6 amp eo o 000000064 G o 0 90 0 00 0 8 9 0 0 e ee 6 OO oO 0 0 Oo oe Oo 20 00 00 00 OO OO OO OO 20 16 0 0 0000 0 OO 0 0 oO 0 o eo o oe eo GG e OG 6 Ll o ic 0080000890000 0 G 8 0 Sooo O G G ogee os 8 o o o 0 0606r00 0 06 0 6G 6 6 o o o ld 6 6 10 0000000010000 0 0000000 0 96 o 0 5 o BF 9 00 0 0 0 0 0 0 0 0 9000009000090 00900 GG G G G OGG 6000600000000 o 0 O o o 8 0 909090008 0 00 009 0 00 00 6 Y 0 0 0 0 90906000000 4 0 0 0 0 8 Ge G6 o o ul GO 00000000 o o o o 0 o e 000060600060 o E E ooo OO Oi Oo 6 20000000 8 o 0 006001006060 00 e o 0000000006 600006060 060
5. it is unwise to use the results of SOLVE_MATRIX_AR in VEC3 analysis SOLVE_MATRIX should be used instead If Type integer equals 1 simple Arnoldi algorithm is being used For the effective performance the maximum number of dipoles should be kept around 10 000 or so maximum Nmodes is a desirable number of modes Nev integer is a performance regulating parameter connected with a number of Krylov vectors used in the simulation Usually the number should be 22 at least 50 but no large than 2 5 Nmodes with an exception of the case when more than 5000 dipoles are being used In this case slightly smaller numbers starting with 30 and reaching maximum 2 Nmodes should be used Typically the number of modes effectively participating in the absorption of uniform r f field does not exceed 25 or so low frequency modes If Type integer equals 2 sparce algorithms will be used with an assumption that the dipole dipole interaction between dipoles separated from each other by the distance larger than Recut double This is a very effective method to calculate the spectra of very large thin film particles Rcut typically can be set to be around 200nm 200D 7 but one needs to check to make sure that the calculations are well converged EIGEN_SAVE_TEXT Type Filename Filename character is the file s name Both the eigenvectors and eigenvalues are going to be saved in a text file Filename Type integer defines whether right Type 1 or l
6. the lowest energy configuration being selected from the results Initial configuration is the same as defined by M_ LOAD or INIT_M for the first run and random for the other runs For each of the runs error and energy are being displayed at the standard output computer screen When error is below maxerror double we recommend the value 1D 5 the process is stopped Magnetization of the sample is written into the file Im dat where I is the run number starting with 1 Magnetic field of the final configuration is written into Ih dat and I wrl contains the VRML picture of the configuration obtained Energies of the runs are written into the file fort 101 Finally the lowest energy configuration is selected among the runs and used as a final equilibrium configuration together with the corresponding magnetic field Koeff double is a coefficient controlling convergence of the method It can have values in between of 0 and 1 with bigger values being recommended 0 8 to 1 0 FFT is used automatically for large samples If koeff is less than zero instead of the relaxation method simple Runge Kutta integration is being performed using the timestep set by SET_TIMESTEP and the damping coefficient defined in SET_BETA Runge Kutta integration is terminated when the difference in energies between two consecutive runs is less than maxerror This method can be more effective when there are strong d c fields present in the system H_SAVE Type Filename
7. c Magnetization dynamics and other related procedures First we need to start with choosing an initial configuration for the magnetization It can be done by loading it from a file M_LOAD Type Filename Type integer is 1 for the text 2 for binary mode Filename character is the file s name Again operating with text files involves certain loss of precision At any time similar command can be used to save the magnetization distribution M_SAVE Type Filename li This command saves the magnetic moments into the file Filename The only difference from M_LOAD is that Type 4 leads to saving of magnetic moments into VRML file Type 1 text file Type 2 binary file VRML scene is created according to the parameters given by SET_VRML The distribution of the magnetization can also be initialized by using INIT_M Type Cx Cy Cz Type integer defines whether the initial configuration is a saturated state Type 2 when all dipoles are pointing in the same direction in this case Cx Cy Cz double define such direction the vector with coordinates Cx Cy Cz does not need to be a unit vector If Type is equal to 1 18 no other parameters need to be entered the initial configuration is defined as a vortex lying in x y plane with the center coinciding with that of the object itself If Type is equal to 3 the initial configuration is random External d c magnetic field is set by using ADD_EXT_F
8. can be used to override the default values of various parameters SET_GYRO Gyro Assigns the value Gyro double to the gyroscopic coefficient y Affects the units in which frequencies and times are being presented For example Gyro 1 75D 2 makes all times to be in nanoseconds and all frequencies in GHz Gyro 2 785D 3 corresponds to using GHz 2x for the frequencies The value by default is 1 SET_BETA Beta Sets the value of damping parameter f to Beta double The value by default is 0 SET_CURRENT Current Cx Cy Cz Describes the presence of a spin polarized current in the system Current double current s strength in Oe 1000 Oe approximately corresponds to 10 A cm Cx Cy Cz double x y and z components of spin polarized current s polarization By default the spin polarized current has a zero magnitude there is no current in the system SET_HRF Type real Cx imag Cx real Cy imag Cy real Cz imag Cz real Cx imag Cx real Cy imag Cy real Cz imag Cz i i2 Filename icur cx cy cz Type integer is equal to 1 for a uniform r f field 2 for a linear field 3 for a field from a file 4 for a field due to the r f spin polarized current The uniform field Type 1 is defined by three complex components Cx Cy Cz complex that are entered as two real numbers each real and imaginary part respectively The rf field itself is then defined as h Re amp C 9C
9. 000 e o G G o o Cs SES Es E Dl e G o DRAW_MODES_P P F I n Instead of printing modes in local coordinates this routine prints projections of the modes with numbers between I and 12 integers onto a global Cartesian coordinate system P integer defines which projection is going to be printed 1 for x 2 for y 3 for z coordinates F double is initial phase The best value to start with is probably 0 but if one suspects traveling spin 27 waves to be present in the system one can try other values from 0 to 2pi to see the modes evolution Example Here you can see the modes in an ellipse Colors denote sign and magnitude of excitations 10 DRAW_MODES E Magnitude Can only be executed after the CALC_C command Prints global and local coordinate representations for all the modes that have absorption higher than Magnitude double maximum absorption I e Magnitue 0 05 means that all of the modes that have absorption strength of at least 5 of maximum will be plotted Different coordinates are distinguished by numbers in the files name 0 for local coordinates from 1 to 3 x y and z coordinates We had a very positive experience using programs like Adobe PhotoShop CS to work with the graphs created by RKMAG END Finishes the program s execution It s appearance at the program s end is mandatory 28 29 Final Notes Can be found at http w
10. 1 SET_SCALE Sx Sy Sz Sets the spatial averaging in the matrix creation as defined in my thesis to that among Sx Sy and Sz integers cells along x y and z axis respectively The use of this command is recommended for only very advanced users The values by default are Sx Sy Sz b Describing the object In our case we do it by using a square lattice composed of cells or discrete dipoles and a masque which defines which lattice sites are occupied and which are not It also tells which material occupies the site Materials are characterized by two parameters exchange stiffness and saturation magnetization First of all we need to input such materials It s done by invoking the command 10 LOAD_MAT N Ai Ms A2 Ms Here N integer is the total number of materials M and A double are saturation magnetization and exchange stiffness for each of the materials If we already have a masque created by previous calculations we can load it by using MASQUE_LOAD Type Filename Type integer tells whether the file filename character is a binary Type 2 text Type 1 or modified text Type 3 file Text file format is as follows I Xmax Y max Z max x y Z masque x y z where I is a total number of elements in the system Xmax Y max Zmax integers are maximum values of x y and z coordinates respectively in number of cells x y z integers are coordinates masque x y z integer is a material
11. APE moran A RE ak eas pas Sa hens ENAA 11 MASO UE CRA TE alas cae as Coates yas e E e a en soso tae 12 ADD SHAPE O 12 MASQUE INFO io A A 14 TENSOR CREATE ti A A RRA 14 ADO OANISOMROPY ao waar E A an alten doe Goats ia Slat scl Gani 15 A D ANISOTROPY O tooo allas 15 TENSOR SAVE srta id ias 16 TENSOR LOAD usais 16 TENSOR MACRO nionane en a a a a a EG TE ee 16 c Magnetization dynamics and other related procedures oooconnccnnnccnononoccnnnnnconncconocann nina ccoo ncnnns 17 MITAD l a 17 M SAVE meoin E EE A EA AEO ea 17 NIE Morie RO 17 ADD EXT FIELD aenn a O aeeous 18 RELAX aeranta AT E ai 18 ESA ME a aa 19 BODA a O A A E a eee 20 TAME e a 20 SET TIMESTEP A A A oases 20 A A O E ers 20 d Solving for eigenvalues and eigenvectors csho 3 cscsca casted educa cca urdahna mi eames 21 SOLVE A Toc cata a ced aaa ade goa Gots aae Oa ude Saniles a oxeab vasa ech eis ao aae Sonne ee 21 SOLVE MAIRI 2A Rita 21 EIGENSAVE TEXT A Ae 22 DONES RN eS 22 EIGEN LOAD Fo ae a tts ne Sed Sauk E aaa 22 CATE Saree cen E EEE E EEEE 22 PRIN O E E E AENEAS 23 PRINT ABSen ative eat E E R EENE EE 23 VEC er E E L amc AE E ON ate eau ak oc ne ae ene caeh 23 E Displaying the Ele desd di salons 24 PRINT PASES iaa 25 PRAV MO DES a E iaa 25 DRAW MODES Peso ios ated ai sao stata Bes ie cat e a Aaa hai dh 26 DRAW MODES Esso AA ERRE 27 Final Notes This program uses Gaussian units with an exception of the units for time and frequency which are b
12. IELD i Cx Cy Cz Filename If i integer is equal to 1 Cx Cy and Cz give the vector components of a uniform external d c field If 1 is equal to 2 the field is read from the file Filename character with j defining the file s type 1 for text 2 for binary It is important to mention that this command can be issued many times with the resulting field being a sum over the contributions created by each of such commands for example some external file can be loaded and than an additional field added by issuing ADD EXT FIELD for the second time INIT_M or M_LOAD should always precede ADD_EXT_ FIELD By default if the command is omitted external field is assumed to be 0 everywhere In order to find an equilibrium configuration one can use so called relaxation or modified Powell method another option is to use Runge Kutta to integrate Landau Lifshitz equation in the presence of damping Both are done by using RELAX I Maxerror Koeff 19 This command finds the equilibrium configuration I integer is a number of configurations that are going to be used If I is equal to 1 only the configuration that was previously defined by INIT_M or M_LOAD is going to be used If I is more than 1 for the first configuration I integer defines how many times the calculation is going to be performed If I is equal to 1 the relaxation is performed only once If I is greater than 1 the relaxation is performed I times with
13. RKMAG USER S MANUAL John B Ketterson K Rivkin Introduction This program is intended to offer an alternative to traditional approaches like using Runge Kutta integration of Landau Lifshitz equation when it comes to calculating resonant spin wave modes in arbitrary shaped bodies The idea behind the method is to formulate the problem of finding the resonant frequencies and resonant modes as an eigenvalue problem Further on upon calculating the resonant modes one can use these modes to calculate the absorption properties of the body in question We would to offer our appreciation to Mr Alexander Heifetz and Dr Paul Sievert for their contribution to this project We also would like to mention that while this method have been developed by us independently we came to know that some of it s elements where previously published in works of P Politi M G Pini A Rettori F C Nortemann R L Stamps R E Campley and their collaborators The program is intended for non commercial use TOC NTO a sce cca ee ahaa cc ay che ee ls tut ee cane a cats E a ac a 2 INSTALO A A iS 6 as Optional COMMANIS ra na en ea E EE O E E E E iaa 7 SET GYRO pena e er a e E Peery ere EE 7 ES 7 SET CURREN T aeann eaa ariga oa a A TERNS 7 SET HRE ind 8 A E a E AE ES 9 SET SCALE nnii e R E E S 9 Bb Describing the Oia da AAA A A E E dat a 9 LOAD MAT init O 10 MASQUE TOA DD a at RE 10 MASQUE SN El EOS 11 O A O E A 11 STUDY SH
14. at Cx Cy Cz K 16 can be used In this case K double gives not the anisotropy coefficient but the value of effective anisotropy field itself in Oe One can save the calculated tensor by using TENSOR_SAVE Mat Type Filename Separate materials should be saved loaded to from different files Mat integer defines which material s tensor must be saved Type integer defines whether the text Type 1 or binary Type 2 should be used Text mode should not be used for permanent storage due to the precision loss but rather only to check the correctness of the tensor Filename character is the name of the file being used for saving into To load the tensor one must use TENSOR_LOAD Mat Type Filename Everything is the same as in the case of TENSOR _ SAVE but now the tensor is being loaded from a file TENSOR_MACRO This command is somewhat unusual it calculates and saves text version into macro txt binary version into macro dat the tensor corresponding to the whole object i e the table that shows the average field inside the occupied sites in the object as function of the object s magnetization In order to avoid huge self fields due to the exchange interaction exchange stiffness in LOAD MAT preferably should be zero This routine can be useful when a macrospin approximation is being desired for a complex system in this case it produces the demagnetization tensor for such a macrospin 17
15. ault can occur Unfortunately current version does not contain check the input dat file for errors Incorrect input data will most likely result in a segmentation fault The program is intended for non commercial use however we do express our desire that all the papers written with the help of RKMAG software packages should acknowledge such fact We also ask you to be so kind as to submit the references to such papers to us by e mailing us at rivkin rkmag com It is also important to mention that the maximum number of individual cells dipoles that one can effectively use with eigenvalue method is 6000 11 000 If one wants to have a better discretization with more dipoles one needs to use one of approximate methods described below for example the best results for thin films can be obtained by using sparce algorithms The program is executed by running the executable file for example rkmag In the same directory there should be present the input data file input dat Input dat should contain commands that define the actions that will be performed by the program The format of each command is COMMAND Parameter 1 Parameter2 Where each new parameter should appear on a new line In this manual parameters that should be entered only for one specific value of other parameters are shown in square brackets Empty lines are thrown away during the input a Optional commands In this section we will describe the commands that
16. bjects Type 1 Rext Rint H double are external internal diameters and the thickness of the disk ring Cx Cy Cz double are coordinates of the ring s or disc s center 13 14 Type 2 Rext double is sphere s diameter Cx Cy Cz double are coordinates of the sphere s center Type 3 Slab is defined as spanning all the space between coordinates x1 and x2 all coordinates are real numbers with double precision by default on x axis yl and y2 on y axis zl and z2 on z axis Type 4 Rext Rext2 H double are semimajor semiminor diameters and the thickness of the disk ring Cx Cy Cz double are coordinates of the ring s or disc s center MASQUE_INFO This command prints the total number of occupied cells in the system followed by the number of cells along x y and z directions After the object s creation one has to calculate the demagnetization tensor This is done by issuing Calculation of the demagnetization tensor is performed by issuing a command TENSOR_CREATE Nx Ny Nz Recut If Nx integer equals 1 the demagnetization tensor is calculated by using Newell s formulas for uniformly magnetized prisms Individual cells in this case are uniformly magnetized prisms and the magnetic fields are macroscopic If Nx gt 0 the demagnetization tensor is calculated by assuming that each individual cells is composed from magnetic dipoles total number Nx Ny Nz where Nx Ny and Nz integers define how ma
17. eft Type 2 eigenvectors are going to be saved In order to print only the eigenfrequencies OMEGA_PRINT Filename Prints all of the resonant frequencies to the file Filename EIGEN_LOAD Reads the eigenmodes from files eigen dat and eigenl dat CALC_C Calculates the oscillation strengths 23 PRINT_C Filename Prints the oscillation strengths to the file Filename character Format is Mode number from 1 to N where N is a number of macrocells real part of oscillator strength imaginary part of oscillator strength real part of eigenfrequency imaginary part of eigenfrequency Imaginary part of oscillator strength should be positive and corresponds to the absorption In case if imaginary part is negative it is most likely due to discretization or some numerical errors unless the initial configuration is not an equilibrium PRINT_ABS Filename Omegastep Prints the absorption graph to the file Filename character omegastep double is the step size for the frequency axis in the units defined by SET GYRO The frequencies chosen for the graph lie mostly in between of the lowest and highest excited modes Absorption is printed in arbitrary units the object s volume is not being taken into account however r f different fields acting on the same object will produce the absorption that are going to be in the same units and therefore can be compared to one another VEC3 Nmodes Endtime Vtime Fname
18. eing defined by the choice of gyroscopic coefficient y see the description of SET GYRO command below The method have been previously described in a number of publications for more information please read the following work K Rivkin and J B Ketterson Micromagnetic simulations of absorption spectra to be published in Journal of Magnetism and Magnetic Materials http www rkmag com articles abs2 pdf Installation Currently the program is intended to work under Linux FreeBSD OS but can be relatively easily recompiled for other operational systems To install the program you need to download the appropriate executable file from http www rkmag com page2 html While the new version of the software is supposed to be stand alone executable it may occur that on certain system the program require certain libraries to be present in the system These libraries are available at www rkmag com page2 html All one needs to do is to place them into one of the following directories lib usr lib usr local lib or other directory that is included in Id search path After doing so ldconfig or etc Idconfig should be executed It s also highly recommended that ulimit s command shall be used before the program s execution in order to set the stack size to a much higher than default value 500 5000 Kilobytes should suffice for most of the cases ulimit s 5000000 would be a safe choice Otherwise the segmentation f
19. eters a b c d Semimajor axis b Semiminor axis a The angle in between of semimajor axis and y axis The initial phase or initial position of an oscillating magnetic moment 25 There are a few routines that allow one to make plots and drawings of the modes PRINT_PHASES I n I and I2 are integers This command prints the information about the modes with numbs in between of I and 12 modes are arranged by numbers from low to high frequencies starting with 1 The print contains X y and z coordinates of a cell in number of cells for example a cell can be identified as 1 cell along x axis 2 cell along z axis and 0 cell along z axis angle formed by an initial position of a magnetic dipole with local y axis angle formed by semimajor axis with local y axis magnitude of semiminor and semimajor radiuses normalized to 1 DRAW_MODES I n I and I2 have the same meaning as in PRINT_PHASES but this time the modes are represented as a two dimensional Encapsulated PostScript EPS graph File s name consists of a 0 followed by dash followed by the mode number from low to high frequencies same number as given by OMEGA PRINT PRINT C or PRINT PHASES Planes with different z coordinate are represented as different slices One has to take in mind that the precessional orbits lie in different individual planes each of them is perpendicular to the local equilibrium magnetic moment In the graphs the ellipses
20. ing supported the following command can be used STUDY_SHAPE Type Rext Rint 12 H Rexe Dx D Rext Rext2 H Where Type integer defines the shape of the object 1 corresponds to a disc ring 2 to a sphere 3 to a slab 4 to an ellipse If Type 1 Rext Rint H double define external internal diameters and the height in centimeters The object lies in x y plane If Type 2 Rex defines the sphere s diameter in centimeters If Type 3 Dx Dy Dz double in centimeters define the length of the slab along x y and z axis If Type 4 Rex is the diameter along the semimajor axis Rex diameter along the seminor axis H is the thickness in centimeters Another way to create a desired object is to use the following two commands MASQUE_CREATE X max Y max Lmax Where Xmax Y max Zmax double are maximum values of x y and z coordinates respectively in centimeters Than one can add shapes by using ADD_SHAPE Type Rext Rint Mat Rext Mat x1 x2 yl y2 zl z2 Mat Rexe Rae Mat This command can be issued multiple times in order to define a sample consisting of many objects Type integer defines the object s shape 1 corresponds to a disc ring 2 to a sphere 3 to a prism 4 to an ellipse All values are given in centimeters In all cases Mat integer is the material for the object Mat 0 corresponds to the empty space and can be used to introduce holes into o
21. ny dipoles are along x y and z axis The field is microscopic Lorentz local field but all physical characteristics resonant frequencies switching times etc should be the same When macroscopic fields are being used 15 the total energy of the system includes certain self energy terms therefore the total energy can be positive Second method is usually faster and in certain cases can be slightly more precise than using Newell s formulas We recommend using something like a dipole per nm for Nx Ny and Nz so that 5x5x25nm cell will have Nx 5 Ny 5 Nz 25 with no less than 5 and no more than 30 for each of Ns Rcut double is a cut off diameter in centimeters above this value instead of calculating the average field per cell only the field at a central point is calculated This diameter should usually be slightly bigger than twice the lattice or discretization constant The Newell s method on the other hand is more straightforward and easier to use In order to add the uniaxial anisotropy into the discussion one must use ADD_ANISOTROPY Mat Cx Cy Cz K Mat integer is the material with the anisotropy Cx Cy Cz double are components of a unit vector that defines the anisotropy s direction K is the anisotropy coefficient Kam A y 2 A This corresponds to adding the effective anisotropy field defined by Hf a H is a unit vector defined by Cx Cy Cz Alternatively ADD_ANISOTROPY_FIELD M
22. occupying the site with such coordinates 0 corresponds to an empty cell positive integers correspond to materials numbered in the same order they are in the LOAD_MAT statement Modified text file Type 3 is a text file with the following format First 4 integers describe the total number of elements zero and non zero in the masque maximum values of x y and z coordinates in the lattice Following them there are depictions of the masque for each of the z coordinates for example z 0 11 0001000 0201002 0002000 in this example we have empty sites denoted by 0 and those occupied by the first and second materials as given in LOAD_MAT by numbers 1 and 2 One can save the masque by using MASQUE_SAVE Type Filename Where the parameters have the same meaning as before Type integer tells whether the file filename character is a binary Type 2 text Type 1 or a modified text Type 3 file only now the file is being saved rather than read For the descriptions of the file formats please see the section on MASQUE_SAVE command If one needs to calculate the demagnetization tensor see later or to use any of the shape defining commands one first needs to use SET_CELL Cellx Celly Cellz Which tells the program the cell s dimensions used cellx celly cellz double in centimeters If one just needs to study a simple body currently rectangular prisms rings ellipses discs and spheres are be
23. volution of the magnetization can be observed The files are numbered starting with 1 1 wrl Besides that the program outputs into the file Filename character time average magnetization along x y and z axis and the system s total energy If r f field is present in the system omega double indicates the r f field s frequency 21 d Solving for eigenvalues and eigenvectors In order to use the procedures described in this section one needs to have 1 Local magnetic moments and magnetic fields 2 Masque 3 Tensor 4 Materials parameters LOAD MAT SOLVE_MATRIX Nmodes Upon completion this command will produce all the eigenvalues and eigenvectors Only the Nmodes integer lowest frequency modes are stored the rest are discarded This is usable since most of the high frequency modes typically do not affect experiments If Nmodes equals 1 all of the modes are stored Right and left eigenvectors eigenvalues and scaling factors as set by SET_SCALE or SET SCALE _ R2 are saved in filed eigen dat and eigenl dat SOLVE_MATRIX_AR Type Nmodes Nev Recut Uses Arnoldi algorithm to calculate selected low frequency eigenvectors and eigenvalues Only the right eigenvectors are being computed with an assumption that they are more or less orthogonal to each other While for the sake of calculating the absorption spectra it is usually a good assumption 5 error in the absorption peaks magnitude is usually tolerable
24. ww rkmag com If you have a question please feel to contact us at rivkin rkmag com If you experience a problem while executing our program please include your input dat file your CPU model and all the output produced by the program this can be obtained by executing rkmag amp gt out txt We greatly appreciate your input
Download Pdf Manuals
Related Search